The International Airline Transport Association pegs the maximum net post-tax profit margin for any region in the year 2019 at a meagre 5.5 % at best. The aviation business is cutthroat with very little margin and heavy competition. Airlines unwilling to change are closing shop. India is quite familiar with this phenomenon and has witnessed the closure of several airlines, starting with Modiluft, Eastwest, Kingfisher and recently, Jet Airways.
The data revolution seems like the most obvious answer to the industry’s woes. Data science has made headway into the airline industry but only in spaces such as ticketing and revenues. Airline direct operating costs are at 50% of total revenue and not much has been done in the area due to a lack of bandwidth and connectivity. But, we are in very interesting times, as SpaceX and other organizations are ready to give the airline industry a boost with low latency gigabit internet connectivity. SpaceX is currently working with the US Department of Defense and has already demonstrated throughput of 610 megabits per second to the cockpit of a U.S. Military C-12 twin-engine turboprop aircraft. In comparison, the fastest internet connection available worldwide on aircrafts is approximately 12 megabit per second and considered a premium service for business jets and a few airliners.
Sign up for your weekly dose of what's up in emerging technology.
Airline operating cost can be loosely categorized into Ownership, Maintenance, Fuel and Crew. The life or airworthiness of an aircraft is measured in pressure cycles endured, meaning a flight from Srinagar to Chennai will cost almost the same as a flight from Mumbai to Pune from an ownership point of view. Dynamically optimizing city pairs, rotation of aircraft, and reducing downtime gives maximum utility and also makes it a challenging data problem that will now have an element of live data streams, making decisions more realistic and long term. Operators with multiple fleets can take advantage of planning based on sales forecasting by interchanging aircraft type depending on the requirement. If an airline has both a Boeing 777 and a Boeing 737, they carry approximately 400 and 200 passengers respectively. The airline could deploy the Boeing 777 if there is an increase in demand for a sector, which is normally operated with a Boeing 737 or vice versa. Time is indeed money for an airline operator and a single aircraft that’s not being utilized optimally can deliver quite a hefty punch on their bottom lines.
Fuel-saving has been identified as the best solution to reduce cost as well as environmental impact. For instance, Airbus considered having almost 10,000 sensors on the wings of its A380-1000 aircraft to try and make the aircraft more fuel-efficient as well as predicts mechanical failures, thus warning the operator beforehand. Traffic-Aware Strategic Aircrew Requests (TASAR) developed by NASA Langley Research Center was selected as an R&D 100 winner, an award for disruptors that will change industries and make the world a better place in the coming years. It works like Google maps and gives information to the pilot regarding which route to fly and what altitudes will be available, thus saving time and fuel. This is based on freely available Automated Dependent Surveillance-Broadcast (ADS-B) data, transmitted by almost all aircraft giving you basic information on position and movement of aircraft, to summarize approximately 13 variables. One can only imagine the scale at which decision making on aircraft will be impacted by complete connectivity and an instantaneous update of aircraft state, weather conditions, estimates of other aircrafts to a common destination. This information can make the utilization of air routes and airports optimal and virtually eliminating most delays. The annual cost of delays was $28 billion in 2018, according to the Federal aviation association’s estimates (direct cost to airlines and passengers, lost demand, and indirect costs). Increasing demands for air travel combined with more frequent weather events due to climate change will only push these numbers higher.
Many an example can be given for catastrophic accidents due to mechanical failures in a single component of an aircraft. Aviation has adopted and managed a very good safety record, but as the scale of operations continue to grow, we need more predictive than reactive solutions. Predictive maintenance has been a solution talked about for the past few decades, but the capability of real time monitoring and analytics is very limited. An Airbus 350 aircraft with about 50,000 sensors onboard collects close to 2.5 Terabyte of data per day. Transferring the data and processing it poses a big challenge with our limited connectivity, which will no longer be the case with reliable internet connectivity, paving the way for usable intelligence where aircraft components can be predictively serviced before failure, reducing delays and highly improving the safety of airliners.
The human factor in airline operations is indispensable and expensive. A job with very little positive feedback, long hours, shift work and haywire schedules. They are indeed paid very well for their services, but monetary gain alone cannot attract the right kind of workforce. There is, equally, a factor of skill compared to knowledge involved in flying an aircraft and people train for a decade or more before they are put in charge of commercial airliners, hence retention of the workforce reduces cost and improves safety. Data science can become the answer to the airline scheduling problem, currently airlines have systems in place to accommodate requests and are given a point-based value for the duties the crew would like to take on, however, managing requests and trying to accommodate everyone can be quite a laborious task. Predictive analytics could be used to roster the crew as well as a trade platform removing the need of human intervention in crew scheduling. Essentially trying to accommodate any request from the airline crew that is possible, considering factors such as legality of operations, qualifications of crew, recurrent training and crew duty time limitations, to name just a few.
Mundane tasks such a monitoring flightpath, fuel and communication during cruise phase for airlines with connectivity could be automated reducing workload and ideally the number of occupants in the cockpit during low workload phases, thereby enhancing safety as well as crew costs per flight. Internet on the aircraft can reduce the cabin crew workload as I don’t see the call bell ringing that often if we can Netflix and chill on our 16-hour flights from New Delhi to San Francisco!!
The IOTA foundation – a blockchain initiative for the Internet of Things – an interesting project by itself, has managed to tokenize data streams of sensors placed by an individual. If I have an air quality sensor, I could sell my data on IOTA’s online platform to a person doing a study on air quality or to a company that needs the data for their app. Spot weather data from an aircraft at 20,000 feet could be invaluable for predictive weather modelling, something that could be tokenized and sold creating value for the struggling airlines. Weather radar data could be passed along to other aircrafts as well as ground stations during more dynamic weather phenomenon, which could lead to the tracking of storms better, saving countless lives as well as billions in revenue. The internet revolution for the airline industry is very nascent and of the petabytes of data being produced, less than 0.5% is being converted into useful intelligence.
The possibilities of creating value from data science in airline operation is limitless. Data revolution might just be the key for getting aviation back on its feet, continuing to connect more people while creating new opportunities.